Spatial Heterogeneity in Tax Sensitivity: Evidence from Cross-State Cigarette Purchases

Differences in excise taxes across states incentivize consumers to make cross-border purchases. In this study, we investigate the “border effect†phenomenon, which refers to the impact of cross-state purchasing behaviors on the excise tax sensitivity of consumption. We analytically formulate the “border effect†as a linear function that decreases with distance from the closest lower-tax state. We reasonably assume that the “border effect†reaches a maximum at the border with the lower-tax state and then linearly decreases with distance from the border, eventually reaching zero after a certain cutoff distance. We estimate the parameters of the “border effect†function employing a threshold regression model with location and time fixed effects. As a robustness check, we also run a segmented regression using separate tax sensitivity estimates for a range of distance intervals. We verify that the estimates from segmented regression align with the linear pattern derived from the threshold model. Further, we enhance the “border effect†function by incorporating a difference between the home state tax and the closest lower-tax state tax as an additional factor, and then compare the estimation results for the two specifications. Beyond geographic variation, we also examine how the tax sensitivity of cigarette consumption differs across demographic groups. Our analysis shows that tax sensitivity varies by income level: high-income consumers are the least responsive to excise tax increases, while low-income consumers are the most responsive, with middle-income consumers falling in between. All income groups are influenced by the “border effect†, but for high-income consumers this effect is only present at distances up to 60 kilometers from the lower-tax state. Our analysis based on employment status indicates that both employed and non-employed consumers display a similar shape in the “border effect†function; however, non-employed consumers show significantly higher tax sensitivity than employed consumers.